By Tom Wolf, PhD, P.Ag.
A wise professor once asked his class what they thought was the most powerful force in the universe. The students produced a lot of answers. “Gravity,” said one. “TNT,” said another. “Nuclear fission,” bested a third. The professor shook his head. “Incremental change. That is the answer,” he said. The students looked at each other. What on earth did he mean?
When charting a course for change, we are inclined to pin our hopes on big things. The breakthroughs. The game-changers. Like a new mode of action to address herbicide resistance. Or a new crop that is easy to grow and offers tremendous returns. And although these do happen from time to time, the majority of change in our lives is stealthier. It comes in small increments, so small that many are hard to even detect. But over time, their impact is impossible to ignore, and even harder to reverse. An example from the banking sector is compound interest. In agriculture, it could be plant breeding. And in spraying, it could well be precision.
Precision application continued to elude spraying when site-specific fertility was already well established. The reason was the inability to map weed populations easily. Spot sprays put an end to that. Suddenly, we can treat just the weeds, and that presents a lot of savings. And it has the potential to be a game-changer. But spot sprays are a large investment, something that may not be right for everyone. And it has risks, such as misses, or gaps in algorithms, where certain crops or weeds are not yet recognized. Is there something more intermediate—more incremental—that leads to the same outcome? Something less flamboyant than spot sprays?
There are several examples that come to mind, and I’d like to highlight two of these.
The first is the technology provided by Croptimistic of Naicam, Sask. (swatmaps.com). Long known for mapping the electrical conductivity (EC) of soils and layering this with elevation maps to produce 10 “SWAT” cropping zones, they’ve upped their offering with weed and crop establishment information. By mounting two cameras on a sprayer (SWAT CAM), one on each boom, they snap photos that capture images of the growing crop. With some artificial intelligence, these images confer information on crop and weed establishment. For crops, information such as plant spacing can teach about the frequency of planter skips and validate the target plant density. It can also teach about the uniformity of plant establishment. Adjustments to seeding, or fertilizer rate, can then be made on a zone-by-zone basis if justified.
Knowledge of the weed distribution across a field is fundamental. We’ve waited a long time for weed density maps so that we can initiate a precision response. Higher weed densities point to the need for more crop competition, more diligent weed control, or both. The traditional way of measuring plant biomass using NDVI (Normalized Difference Vegetation Index) doesn’t discriminate between crop and weed plants. The AI employed on the images does. This information isn’t meant to trigger individual nozzles. Instead, it strengthens the fundamentals of elevation and soil characteristics and applies them to crop and weed growth.
If weed establishment can be tracked, sprayer settings can be managed. This would be done on a whole boom basis, with changes in dose implemented via spray pressure or travel speed. For some sprayers equipped with direct injection, it may be possible to adjust herbicide rates or tank mixes. Of course, the recurring appearance of high weed densities in some zones justifies a change of strategy that goes beyond herbicide application.
Fungicide application can also be matched to SWAT zones that represent high or reduced yield potential based on crop images. Yes, SWAT CAM is another layer of information that has to be managed. But it’s a step towards greater precision, accuracy, and product stewardship.
A second Canadian venture is Geco Strategic Weed Management (geco-ag.com). The company was founded to help manage weed populations by attempting to predict where they will emerge. Instead of using boom-mounted cameras, Geco uses historic satellite images to identify field areas with recurring weed patches. It requires a cropping history dating back five years, herbicide history dating back two years, cropping intention, as well as soil type information from the grower. Although satellite imagery is much coarser than close-up photography, it has the advantage of making dozens of images per year available, allowing crop progress to be tracked. Its algorithms can separate crops from weeds in a temporal way.
A subscriber receives three pieces of information. First, a historical trend of the weedy patches in their fields. Second, an indication of patches with potential resistance. And finally, a prescription for where Geco estimates the weeds will occur the next season.
The Geco system is meant to be complementary to other, more finely resolved approaches. For example, their prescriptions could be used to apply residual products. During the growing season, this would help reduce the pressure on a post-emergent spray or allow a spot spray to produce greater savings. Or it could inform seeding rates.
When it comes to change, the size of the step isn’t nearly as important as the direction.
Three prescription maps are produced for the customer, with three levels of conservatism represented by projected hit rates from 85 to 95 per cent. The higher hit rate map covers more area to be treated, hitting more of the weed patch and using more herbicide. The 85 per cent hit rate would save more product and emphasize just the hot spots. The individual user would decide where their comfort zone is.
As with all plant sensing technologies, there will be some uncertainty. And that’s where Geco has identified a matrix of scenarios that can help users. It’s the classic 2×2 matrix. On the x-axis, we have two weed situations: weedy or weed-free. On the y-axis are two responses: sprayed or unsprayed.
The four scenarios are then populated with outcomes. If no weeds are present and the region is unsprayed, it’s a true negative and there is no cost. If weeds are present and they are sprayed, it’s a true positive, and the cost is the treatment, but the benefit is the yield protection. If there are no weeds and the area is sprayed, that’s a false positive. An example would be having a small patch but spraying a large area; the perimeter would be treated with little prospect for a benefit, and could represent waste. And the final scenario is where there are weeds, but the region is unsprayed. That’s a false negative, and it’s costly.
A user would evaluate their comfort zone for each scenario and choose the level of conservatism that matches their temperament.
Geco says their clients don’t need any special equipment. They submit a request and the required information and get a report and shape files for their fields.
Both of these examples represent an intermediate size step to greater precision and accuracy. They also add something that spot spraying isn’t yet doing widely—incorporating the findings into longer-term agronomic strategies and monitoring the results.
When it comes to change, the size of the step isn’t nearly as important as the direction. And although we always want beneficial change to happen quickly, biological systems tend to respond best to small, but strategic change over time. For those reasons, it’s worth looking at these incremental technologies.